IM @ BioHPC

IM (by Jody Hey and Rasmus Nielsen) is an implementation of the MCMC method for the analysis of genetic data under the Isolation with Migration model of population divergence. IM applies this model to genetic data drawn from a pair of closely related populations or species. The results are estimates of the marginal posterior probability densities for each of the model parameters. For detailed description of the IM program and for literature references please consult the IM web page.

All new runs will be submitted with the latest version of IM (revision of 3/5/07). If you wish to continue your calculation from a checkpoint file generated with previous version of the program (revision of 7/31/06), select the appropriate "Program version" button (it will appear once "restart" is selected as the run type). Please note that the current version of the program will not work with checkpoint files generated by the previous version!

Calculations will be carried out on the BioHPC compute cluster at CBSU. You will receive e-mail notifications when the job is submitted, when it starts, and when it is finished. Output will be available via links embedded in the notification e-mails. For more information about this program and BioHPC interface in general, please visit our Frequently Asked Questions page.

E-mail: (only guests need to use this field, registered users should log in)

Job name: (will be set to the name of input file)

Run type new restart

Output file: (will be set after input file is selected)

Comment (optional): (no spaces please)

For full description of all available program options refer to IM Documentation. If you intend to use the "-p" option, read this important information.

Options (write in one line - text will wrap around; options -i, -o, -e, and -h will be included automatically).
Please DO NOT specify very long burn-in times with "-b" option. During burn-in the checkpoint file is not written, so if your job times out before burn-in is over, you will not be able to restart it. The best way to achieve a long burn-in is to combine it out of several shorter runs, each restarted from the checkpoint file produced by the previous run with the option "treat previous run as burn-in" checked.

Cluster: ( Show timeout info )